[FFmpeg-user] Compiling FFmpeg for 32 bits (i386) on Mac OS X

[Michael Bradshaw]

On Fri, Jun 1, 2012 at 11:01 AM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> Michael Bradshaw <mbradshaw <at> sorensonmedia.com> writes:
>
>> libswscale/x86/rgb2rgb_template.c: In function ‘rgb15tobgr24_MMX’:
>> libswscale/x86/rgb2rgb_template.c:814: error: can't find a register in
>> class ‘GENERAL_REGS’ while reloading ‘asm’
>
> It seems the problem is old:
> https://gist.github.com/1528236
>
> I don't have access to an Intel-Mac, I suspect you can disable
> rgb15tobgr24_MMX for the specific gcc version on OS X.

I think that may need to be done. --disable-asm works, but I'd rather
enable assembly.

> Does it work with clang?

No, it runs out of registers too, but on h264_cabac.c.

Configuration:
./configure --enable-shared --disable-static --cc="clang -m32"

make V=1:
clang -m32 -I. -I./ -D_ISOC99_SOURCE -D_FILE_OFFSET_BITS=64
-D_LARGEFILE_SOURCE -DPIC -DHAVE_AV_CONFIG_H -std=c99
-fomit-frame-pointer -fPIC -g -Wdeclaration-after-statement -Wall
-Wno-parentheses -Wno-switch -Wno-format-zero-length
-Wdisabled-optimization -Wpointer-arith -Wredundant-decls
-Wno-pointer-sign -Wcast-qual -Wwrite-strings -Wtype-limits -Wundef
-Wmissing-prototypes -Wno-pointer-to-int-cast -Wstrict-prototypes -O3
-fno-math-errno -fno-signed-zeros -mllvm -stack-alignment=16
-Qunused-arguments  -MMD -c -o libavcodec/h264_cabac.o
libavcodec/h264_cabac.c
fatal error: error in backend: Ran out of registers during register allocation!
make: *** [libavcodec/h264_cabac.o] Error 1